-
Mobile Robot Positioning System of Adaptive Unscented Kalman Filter with Forgetting Factor
Wenliang Zhu,
Junjie Huang,
Yunpeng Zhou,
Chengxiao Zhu
Issue:
Volume 7, Issue 6, November 2022
Pages:
106-112
Received:
6 November 2022
Accepted:
21 November 2022
Published:
30 November 2022
DOI:
10.11648/j.mcs.20220706.11
Downloads:
Views:
Abstract: For the problem of inaccurate positioning of mobile robots in complex industrial environments, a multi-sensor combination localization method for omnidirectional mobile robots is proposed that incorporates the unscented Kalman filter (UKF), Real-Time Kinematic (RTK), and Inertial Measurement Unit (IMU). Firstly, the position information of the mobile robot is obtained by Real-Time Kinematic (RTK) and Wheel Odometry respectively. Secondly, the inertial measurement unit (IMU) determines the cart yaw angle while dual RTK is proposed to solve the yaw angle in real-time. Finally, the position and yaw angle data are input to the unscented Kalman filter in real time. This paper proposes the F-AUKF algorithm, which optimizes the traditional unscented Kalman filter algorithm by introducing a forgetting factor in order to improve the robustness of mobile robot localization for continuous operation in complex industrial building environments. The experimental results show that the F-AUKF algorithm eventually achieves a positioning accuracy about 10 times higher than that of a single odometer, about 6 times higher than that of a single RTK and about 3 times higher than that of the traditional UKF algorithm, effectively improving the problem of dispersion of the filtering effect after a long period of operation and providing better stability.
Abstract: For the problem of inaccurate positioning of mobile robots in complex industrial environments, a multi-sensor combination localization method for omnidirectional mobile robots is proposed that incorporates the unscented Kalman filter (UKF), Real-Time Kinematic (RTK), and Inertial Measurement Unit (IMU). Firstly, the position information of the mobi...
Show More
-
An Alternative Method of Floyd-Warshall Algorithm Is Proposed for Solving Shortest Path Problems Using Triangular Procedure
Md Zahidul Islam,
Md Asadujjaman,
Mahabub Rahman
Issue:
Volume 7, Issue 6, November 2022
Pages:
113-117
Received:
25 October 2022
Accepted:
18 November 2022
Published:
8 December 2022
DOI:
10.11648/j.mcs.20220706.12
Downloads:
Views:
Abstract: This paper presents a new approach for finding the shortest path compared with the Floyd-Warshall algorithm which is mostly used for determining the shortest path between every pair of nodes of network modeling problems. Finding the smallest route through a road network is one of the innumerable real-world applications of the shortest path problem. In this paper, we will discuss the existing Floyd-Warshall algorithm. Then we will explain our new approach to solving shortest-path problems. The method is developed based on a right-angle triangle. This technique of solving the shortest path problem brings out the same result as the existing Floyd-Warshall algorithm. Additionally, we demonstrate that the foundation of our algorithm is simpler to comprehend, which could be helpful for instructional reasons. An example verifies our algorithm and demonstrates how it is used. We hope this paper will give the reader an idea of the network modeling problems and their efficacy, enumerate the benefits gained, and identify areas for further improvement.
Abstract: This paper presents a new approach for finding the shortest path compared with the Floyd-Warshall algorithm which is mostly used for determining the shortest path between every pair of nodes of network modeling problems. Finding the smallest route through a road network is one of the innumerable real-world applications of the shortest path problem....
Show More
-
AGV Positioning Based on Multi-sensor Data Fusion
Wengliang Zhu,
Yunpeng Zhou,
Junjie Huang,
Shukai Guo
Issue:
Volume 7, Issue 6, November 2022
Pages:
118-123
Received:
21 November 2022
Accepted:
5 December 2022
Published:
15 December 2022
DOI:
10.11648/j.mcs.20220706.13
Downloads:
Views:
Abstract: In recent years, with the rapid development of robot technology and explosive growth of robot demand, AGV robot has gradually infiltrated into many aspects of human production and life, and has become a global hot research direction. However, due to the changeable and compact working environment, AGV robot still has many technical problems to be solved. The localization of AGV robot is the premise and key for AGV robot to move freely. To address the problem of accumulated error in wheel Odometry positioning and data drift in ultra-wideband (UWB) positioning when positioning AGV robots in an unknown environment, this paper establishes the coordinate system of AGV robots based on an independently built AGV robot motion control system, and combines the advantages and disadvantages of wheel Odometry and UWB positioning sensors, and uses the TEKF algorithm to fuse the positioning data of the two sensors The TEKF algorithm is used to fuse the positioning data of the two sensors in order to improve the positioning accuracy of the AGV robot. The experimental results show that the integrated positioning system of wheel Odometry and UWB can effectively restrain the cumulative error and data drift, and the positioning accuracy of multi-sensor fusion positioning is greatly improved compared with that of a single sensor, providing accurate and reliable positioning data for the motion control of AGV robot.
Abstract: In recent years, with the rapid development of robot technology and explosive growth of robot demand, AGV robot has gradually infiltrated into many aspects of human production and life, and has become a global hot research direction. However, due to the changeable and compact working environment, AGV robot still has many technical problems to be so...
Show More
-
Application Research of Graph Neural Networks in the Financial Risk Control
Zhongbao Yu,
Jiaqi Zhang,
Xin Qi,
Chao Chen
Issue:
Volume 7, Issue 6, November 2022
Pages:
124-129
Received:
8 April 2022
Accepted:
26 December 2022
Published:
28 December 2022
DOI:
10.11648/j.mcs.20220706.14
Downloads:
Views:
Abstract: Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.
Abstract: Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an im...
Show More
-
Application of Dynamic Programming to Revenue Management: The Optimum Validity Model’s Test (S)
Felix Obi Ohanuba,
Everestus Okafor Ossai,
Precious Ndidiamaka Ezra,
Martin Nnaemeka Eze
Issue:
Volume 7, Issue 6, November 2022
Pages:
130-143
Received:
4 December 2022
Accepted:
19 December 2022
Published:
29 December 2022
DOI:
10.11648/j.mcs.20220706.15
Downloads:
Views:
Abstract: A suitable decision plan is followed by an effective financial management to achieve optimality while investing in a competing stock portfolio. This study altered a Dynamic Programming (DP) model of Bellman. The modified model was used to solve a business problem. The problems of choosing a stock portfolio for optimal return among investors in financial markets have resulted in a financial crisis. Most financial analysts provide investors with incorrect and unvalidated investment information. The consequences were minimal optimum, no return, and an investment problem. The goals are to ensure optimality in investor returns, validate the results using two validity tests, and select the test that best validated the model. The silhouette and Dunn tests were used to validate the outcome result. The results of using Silhouette reduced computational complexity and produced a more robust and validated return. The k-means clustering (an aspect of unsupervised machine learning) provided better statistical evaluation and information on the investment pattern. In comparison to previous work, the introduction of variables allowed for the best return at stage one. Finally, a validated investment report can help to avoid mistakes made by market analysts and investors when making investment decisions.
Abstract: A suitable decision plan is followed by an effective financial management to achieve optimality while investing in a competing stock portfolio. This study altered a Dynamic Programming (DP) model of Bellman. The modified model was used to solve a business problem. The problems of choosing a stock portfolio for optimal return among investors in fina...
Show More
